LIMITS OF HEREDITARY CONTROL 867 
‘In grouping the variates shown in the table for purposes of seria- 
tion we have decided to fix the group size on as logical a basis as 
possible. The average range of variability of the sets of quad- 
ruplets is eight scutes. We shall therefore seriate the array in 
groups of eight for reasons which will become clear later. The 
polygon of variation obtained by this grouping represents a close 
120 
110 
100 
90 
80 
70 
60 
50 
40 
+ N oO © o + N (=) co oO s N So wo 
N oa) Tt +t w o ~ ioe) oO fon) fo) -_ N N 
wo w wo nn w w w wo wo wo ive) Ke} oO o 
' | ! 1 ' | ! 1 ! ) ' ' ! 1 
~ uw) oO - fon) ~ w on oO (op) > wo oo) - 
— N ~ wo ~ foo} (ep) fo) _ N 
ayy a | a «a * hats Bb wm tS 6 6 
Fig. 1 Polygon of variation for the total number of scutes in the nine bands, 
as determined from a seriation of 508 individuals. Class range = 8 scutes. The 
solid line represents the observational and the broken line the theoretical normal 
curve. In this and the succeeding figure the abscissae refer to number of scutes 
and the ordinates to the number of individuals. 
approximation to a normal array as may be seen from a compar- 
ison of the observed and the calculated curve (fig. 1). 
The three important constants of the frequency polygon prac- 
tically coincide; the median being 558; the mean, 558.2; and the 
mode, calculated from these two constants, 557.6. On the ex- 
